RVPU · configurable robotics vision
Robotics Vision, Deployed in a Day
The RVPU — Robotics Vision Processing Unit — ships pre-compiled, configurable computer-vision pipelines on AMD Kria system-on-modules. Select a pipeline — stereo depth, object detection, EO/IR tracking, gimbal-stabilized detection — point it at your camera, and outputs land in ROS 2 and MAVLink. Pipelines, not models. App config, not model code. No in-house AI or FPGA team required.
TerraBot X
Configurable vision pipelines · ROS 2 / MAVLink native · on-platform
Pre-compiled vision pipelines · in-mission bitstream switching · 8,192-entry cycle-accurate hardware trace buffer
Demo target
30+ Hz
Depth Anything V2 · Kria KV260 · preliminary
Integration
ROS 2
and MAVLink, native
Deploy
1 day
not a six-month AI program
Trace
8,192
cycle-accurate trace entries
Designed for
Qualification on roadmap · reports to design partners under NDA
MIL-STD-810H
Vibration + shock — designed for, qualification on roadmap
−40 to +85 °C
Target operating range
Ruggedized
Ingress-protected enclosure (target)
ROS 2 · MAVLink
Native autonomy integration
Trace buffer
8,192-entry cycle-accurate · forensic observability
Secure boot
Hardware root of trust · per-module attestation
Designed for
The buyers who can't ship on commercial silicon
- Vertical detail
Aerial robotics & drones
Obstacle avoidance, EO/IR & tracking pipelines
- Vertical detail
Ground robotics
SLAM assist, detection & sensor-fusion pipelines
Defense ISR
Sovereign, resilient vision at the tactical edge
One RVPU · two pipeline sets
Same appliance. Configured for the platform you fly or drive.
Both SKUs are the same RVPU appliance on an AMD Kria KV260 system-on-module. They differ by target environment and pipeline set — aerial vision for AeroScale V1, ground-robotics vision for TerraBot X.
RVPU · aerial · Kria KV260
AeroScale V1
RVPU vision appliance for aerial platforms — obstacle avoidance and stereo depth, EO/IR detection, gimbal-stabilized tracking, and BVLOS scene understanding on an AMD Kria KV260.
- Operating range
- −40 to +85 °C (target)
- Pipelines
- Aerial vision · obstacle avoidance, EO/IR, gimbal tracking, BVLOS
RVPU · ground robotics · Kria KV260
TerraBot X
RVPU vision appliance for ground robotics — visual SLAM assistance, person/object detection, sensor fusion, and AMR scene parsing on an AMD Kria KV260.
- Operating range
- −40 to +85 °C (target)
- Pipelines
- Ground robotics · SLAM assist, detection, sensor fusion, AMR scene parsing
Why the RVPU
Deploy robotics vision in a day — not a six-month AI program.
Pipelines, not models
The RVPU ships pre-compiled, configurable computer-vision pipelines — stereo depth, object detection, EO/IR tracking, gimbal-stabilized detection. Select a pipeline and point it at your camera. No model code, no AI engineering.
No AI or FPGA team required
App config, not model code. The RVPU eliminates the need for an in-house AI or FPGA team — teams that would take six months to stand up perception deploy working vision in a day.
Vision-specific by design
Built for robotics perception, not repurposed from a general-purpose accelerator. Bitstream switching lets a platform reconfigure its pipeline mid-mission for the task in front of it.
ROS 2 & MAVLink native
Pipeline outputs land directly in your robotics and flight stacks. First-class ROS 2 and MAVLink integration means the RVPU drops into an existing autonomy pipeline without glue code.
Sovereign & resilient
Defense-grade and built for sovereign supply — deployable under any country or region’s sovereignty requirements, with no dependence on a foreign cloud or data path. Perception stays on-platform and keeps working in contested, comms-denied environments.
Secure by design, forensic by default
Hardware root of trust, signed firmware, and per-module attestation keep the device tamper-evident. An 8,192-entry cycle-accurate hardware trace buffer gives forensic-grade observability for mission review and ROE compliance.
Invotet SDK
Configure a pipeline. Deploy it to the RVPU.
A Python SDK for configuring and deploying pre-compiled vision pipelines to the RVPU — describe your camera, pick a pipeline, and stream outputs to ROS 2 and MAVLink. It ingests PyTorch, ONNX, and HuggingFace models when a pipeline needs a custom detector, with no CUDA in the loop. App config, not model code.
Framework
PyTorch
Bring a custom detector: trace or torch.export models fold into a pipeline with no rewrite.
Framework
ONNX
Standards-based interchange — any ONNX-exported model can back a pipeline stage.
Framework
HuggingFace
Vision checkpoints load through a one-line loader when a pipeline is customised.
For developers
Real entry points, not a wall of marketing
- Open
Documentation
API reference, runtime, and module integration guides
- Open
Quickstart
From a HuggingFace checkpoint to a deployed module in minutes
- Open
Model Explorer
Tested checkpoints across LLM, VLM, and perception families
- Coming soon
GitHub
Open-source examples, model recipes, and SDK source
Bring your camera and your autonomy stack.
The RVPU is available through our design-partner program. Tell us the platform, the pipelines you need, and the integration target — we will set up an evaluation and share the right documentation, including qualification reports under NDA as they complete.